High performance readout circuits and devices for lorentz force resonant cmos-mems magnetic sensors
- Autores: José Maria Sánchez Chiva
- Directores de la Tesis: Jordi Madrenas Boadas (codir. tes.), Daniel Fernández Martínez (dir. tes.)
- Lectura: En la Universitat Politècnica de Catalunya (UPC) ( España ) en 2020
- Idioma: español
- Tribunal Calificador de la Tesis: Maria Aranzazu Uranga del Monte (presid.), Jordi Cosp Vilella (secret.), Francesc Serra Graells (voc.)
- Programa de doctorado: Programa de Doctorado en Ingeniería Electrónica por la Universidad de las Illes Balears y la Universidad Politécnica de Catalunya
- Materias:
- Texto completo no disponible (Saber más ...)
- Resumen
In the last decades, sensing capabilities of martphones have greatly improved since the early mobile phones of the 90’s. Moreover, wearables and the automotive industry require increasing electronics and sensing sophistication. In such echnological advance, Micro Electro Mechanical Systems (MEMS) have played an important role as accelerometers and gyroscopes were the first sensors based on MEMS technology massively introduced in the market. In contrast, it still does not exist a commercial MEMS-based compass, even though Lorentz force MEMS magnetometers were first proposed in the late 90’s.
Currently, Lorentz force MEMS magnetometers have been under the spotlight as they can offer an integrated solution to nowadays sensing power. As a consequence, great advances have been achieved, but various bottlenecks limit the introduction of Lorentz force MEMS compasses in the market. First, current MEMS magnetometers require high current consumption and high biasing voltages to achieve good sensitivities. Moreover, even though devices with excellent performance and sophistication are found in the literature, there is still a lack of research on the readout electronic circuits, specially in the digital signal processing, and closed loop control. Second, most research outcomes rely on custom MEMS fabrication rocesses to manufacture the devices. This is the same approach followed in current commercial MEMS, but it requires different fabrication processes for the electronics and the MEMS. As a consequence, manufacturing cost is high and sensor performance is affected by the MEMS-electronics interface parasitics.
This dissertation presents potential solutions to these issues in order to pave the road to the commercialization of Lorentz force MEMS compasses. First, a complete closed loop, digitally controlled readout system is proposed. The readout circuitry, implemented with off-the-shelf commercial components, and the digital control, on an FPGA, are proposed as a proof of concept of the feasibility, and potential benefits, of such architecture. The proposed system has a measured noise of 550 nT / vHz while the MEMS is biased with 300 µA rms and V = 1 V . Second, various CMOS-MEMS magnetometers have been designed using the BEOL part of the TSMC and SMIC 180 nm standard CMOS processes, and wet and vapor etched. The devices measurement and characterisation is used to analyse the benefits and drawbacks of each design as well as releasing process. Doing so, a high volume manufacturing viability can be performed. Yield values as high as 86% have been obtained for one device manufactured in a SMIC 180 nm full wafer run, having a sensitivity of 2.82 fA/µT · mA and quality factor Q = 7.29 at ambient pressure. While a device manufactured in TSMC 180 nm has Q = 634.5 and a sensitivity of 20.26 fA/µT ·mA at 1 mbar and V = 1 V . Finally, an integrated circuit has been designed that contains all the critical blocks to perform the MEMS signal readout. The MEMS and the electronics have been manufactured using the same die area and standard TSMC 180 nm process in order to reduce parasitics and improve noise and current consumption. Simulations show that a resolution of 8.23 µT /mA for V = 1 V and BW = 10 Hz can be achieved with the designed device.
